A White Paper Best Practices in Automated Agentless IT Monitoring
Best Practices in Automated Agentless IT Monitoring Introduction Agentless IT monitoring is growing in popularity because of its potential to save organizations time and money compared with agent-based products. These agentless monitoring products share a signature attribute: They remove the requirement to install software agents on the systems being monitored. This leads to immediate deployment, lower software license fees, and streamlined operation compared with agent-based alternatives. Without agents to update and maintain, the reduced effort involved can cut the total cost of ownership (TCO) associated with an application or server. To achieve the goals of being affordable, simple, and lightweight, agentless products began as lite monitoring. Instead of being defined by what they did, they were defined largely by what they did not do. This has changed as industry-standard and open source technologies have evolved to permit improved data collection without agents. While today s most advanced agentless monitoring overcomes the limitations of the past, it s important to make a well-informed decision. Agentless Monitoring Then Solutions on the market still consist mainly of older agentless products and tools that lack robustness. Single operating system support Very limited application coverage What should you expect from next-generation Need for add-on layered software agentless monitoring? A best practices approach No corrective action options would suggest a solution that blends the most desirable aspects of agentless and agent-based Inflexible approaches. This paper draws upon Heroix Raw statistics without cohesive view Longitude software as an example showing that it Very limited reporting is possible for agentless monitoring to offer simplicity and savings along with full-featured capabilities suited to enterprises large and small. Agentless Monitoring Now Best Practices Agentless Single OS support Limited application coverage Inflexible Raw statistics Limited reporting Lower cost Agentless Easy to use Immediate deployment Immediate ROI Heroix Longitude: Best of Both No layered software required Multiplatform Comprehensive application coverage Automated corrective action Customizable SLA monitoring Rich reports & graphs Agent- Based High cost of license & implementation Agents required everywhere Steep learning curve Long deployment Shelfware
Then: Single Operating System Support Now: Multiplatform Multiplatform support provides a significant advantage when the varied applications in your enterprise run on different types of servers, and you need consistent monitoring capabilities across them all. Page 3 Some existing agentless products have been adapted from legacy agent-based solutions originally tied to one operating system. Others consist of, or have roots in, shareware. In both cases this led to agentless products developed for use with a single platform such as the Windows or Linux operating system. Starting from square one, Heroix engineered Longitude as a multiplatform solution. Longitude supports six different platforms: Windows, Red Hat, SuSE, Solaris, HP- UX, and AIX operating systems. Support also encompasses a broad range of Web servers, relational databases, J2EE application servers, and email/messaging servers. Then: Very Limited Application Coverage Now: Extensive Application Coverage Preconfigured Without software agents residing on the monitored systems, agentless monitoring has generally been limited to collecting basic statistics. With few parameters to draw upon, this form of monitoring could do little to characterize the performance or availability of any given application. The trend now is to build on more robust industry-standard and open source application program interfaces (APIs), such as SSH, Telnet, WMI, rexec, HTTP, JDBC, and JMX, to endow agentless monitoring with capabilities that measure up to much more expensive agent-based software. Heroix Longitude monitors a range of applications out of the box, including Web, database, J2EE, and messaging servers; infrastructure components; common transactions; and usage trends. Longitude Prepackaged Monitoring Solutions Web Servers Databases J2EE Application Servers Microsoft IIS Apache Web Server Microsoft SQL Server Oracle MySQL IBM WebSphere BEA WebLogic JBoss Messaging Infrastructure User & Business Metrics Microsoft Exchange Server 2000/2003 Active Directory Cisco Devices Dell OpenManage DHCP Transactions Usage Trends Service Level Agreements HP Systems Insight Manager SNMP devices
Page 4 Preset default thresholds take the guesswork out of getting started. It s also easy to refine what is monitored to match your business processes. When in use, Longitude compiles and presents minimum, maximum, and average values reflecting your environment in action so that you can adjust monitoring thresholds based on up-to-date information. Adjustable Default Thresholds Preset application monitoring thresholds speed deployment of Heroix Longitude, which can be done in as little as 15 minutes. You can use your environment s minimum, maximum, and average values to tune the solution over time to reflect characteristics particular to your situation.
Page 5 Then: Need for Add-On Layered Software Now: Self-Contained To keep costs down, the typical agentless product has provided extremely limited functionality. Often you would have to purchase separate software, such as database and report writer packages, and layer it over the monitoring product. The requirement to purchase and maintain additional software drives up TCO, which reduces the savings you would expect from an agentless approach. In contrast, Longitude is self-contained with no prerequisites for layered software. The solution further reduces cost and overhead by being self-contained in another sense: Longitude offers the option to download its own software updates automatically, and handles its own database administration and other internal maintenance. Then: No Corrective Action Options Now: Automated Corrective Action As the name suggests, until recently agentless monitoring has been confined to basic monitoring you had no options to invoke corrective action and generate meaningful alerts. The raw metrics provided did not lend themselves to reducing the time needed to resolve issues, nor did they help you stop minor events from spiraling into major disruptions. Longitude enables email notification, SNMP trap notification, and automated corrective actions. With one screen you can define desired notifications and actions for a single event or hundreds of events. For example, you might instruct the software to restart any autostart service that is not running or email the Oracle DBA when any critical Oracle event occurs. Then: Inflexible Now: Customizes Easily The streamlined nature of older agentless monitoring tools also limited their flexibility. When customization to your specific business environment was called for, you had to have the know-how and time to program scripts. A next-generation product like Longitude demonstrates that an agentless approach can accommodate easy customization. For example, its Event Correlation engine helps capture the business impact of technical issues by combining events across different applications, databases, servers, and platforms from end-to-end. Longitude enables you to detect situations unique to the business services your IT resources deliver and to specify instructions for notifications and actions, all by making selections from a Web-based interface. In addition, role-based security allows you to set up user privileges with specific views that control what people see and access. It may be appropriate to present different views to the network manager and the database administrator, for instance.
Page 6 Customization Using Web Interface For instance, using the Web-based interface you can define a correlated event that will alert you when a combination of events affects a critical business process. Here, a customer-facing order entry system could be compromised by a problem with the application s Web interface, back-end database, or network connectivity. Then: Raw Statistics Now: View of IT Performance and Service Delivery Capable of collecting limited metrics, the first agentless products were not intended to aggregate and present information about IT performance and service delivery. You were left on your own to find meaning in the deluge of data. That was true whether you needed to understand the behavior of multiple resources involved in delivering a business service, or to document SLA performance. The lack of a cohesive view into IT performance and service delivery is increasingly unacceptable to most organizations. The Event Correlation feature (discussed in the previous section) is one way that Heroix Longitude turns data that is monitored into useful information. In addition, more IT departments are being called upon to document SLA compliance whether formally or informally. SLA features integrated into Heroix Longitude translate technical metrics to a good/degraded/unacceptable assessment of service delivery; this supplies a clear picture of how well IT is performing against stated objectives. The
Page 7 presentation of the data reflects maintenance schedules, clustered resources, and services that span applications, servers, and platforms. SLA information is available from both a real-time dashboard and historical reports. SLA Dashboard Next-generation agentless monitoring can provide a dashboard showing the health of all IT resources that contribute to delivering a business service, and indicating whether that service is being compromised in any way. This Enterprise Resource Planning application depends on the availability of a Web page, network performance, memory utilization, database availability, and CPU performance. Then: Very Limited Reporting Now: Preconfigured and Customizable Reports and Graphs Reporting has been an extra for typical agentless monitoring tools. Preparing reports and graphs often required add-on software, which diminished the savings that agentless tools were designed to provide. Agentless monitoring can now be designed to take the effort out of reporting, including on-demand analyses as well as batch reports that can be scheduled to run automatically and emailed to recipients on a regular basis. Preconfigured, customizable reports and graphs can cut many hours from reporting duties. Longitude, for example, comes with hundreds of packaged, customizable reports that can document SLA compliance, match IT performance to business requirements, help analyze developing trends, forecast server capacity needs, and more.
Page 8 Ad hoc and Scheduled Reports Longitude generates reports for all levels of management and technical use. Capabilities include ad hoc reports as well as regularly scheduled reports distributed by email. Above, a dashboard shows an overview of Windows performance (Windows Overview) and provides drill-down to more detailed graphs as well as supporting data (Memory Summary).
Page 9 Conclusion Agentless monitoring is evolving beyond its origins: low-cost solutions that either lengthened the life of legacy monitoring products or took advantage of shareware. Cost, rather than capability, was their focus. Because the market continues to abound with earlier-generation and less-capable products, getting the features and benefits you need still requires an informed look at the solutions you are considering. A purpose-built agentless solution, such as Heroix Longitude, supports a best practices monitoring approach by using industry-standard and open source technologies to offer economical, easy-to-use monitoring with rich functionality once obtainable only from expensive, complicated agent-based products. About Heroix Heroix delivers award-winning software products that help organizations optimize the availability and performance of applications, databases, systems, and IT infrastructure running across multiple platforms. Today, with a presence in over 40 countries, Heroix has delivered performance management solutions to over 1,000 organizations worldwide, including Arrowsight, Fallon Clinic, Kaiser Permanente, Honeywell, Tembec, Motorola, Inc., and Young Broadcasting. Heroix believes that the information in this document is accurate as of its publication date; such information is subject to change without notice. Heroix is not responsible for any inadvertent errors. Heroix, Longitude, RoboMon, Heroix eq, and their corresponding logos are registered trademarks of Heroix. All other company and product names mentioned are used only for identification purposes and may be trademarks or registered trademarks of their respective companies. Copyright 2007 Heroix. All rights reserved.